Research map / Multi-agent systems and games / Cooperative multi-agent RL
Deep Coordination Graphs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add pairwise payoff functions and message passing to represent coordinated actions | VDN-style value factorization · Relative overgeneralization predator-prey (p=-2) | greedy test episode return | Helped |
| Add pairwise payoff functions and message passing | decentralized value-factorization baselines · Artificial decentralization predator-prey | greedy test episode return | Helped |
| Use fully connected rather than CYCLE, LINE, or STAR coordination topology (ablation) | DCG · Relative overgeneralization predator-prey | greedy test episode return | Helped |
| Remove parameter sharing between utility and payoff functions (ablation) | DCG · Relative overgeneralization predator-prey | greedy test episode return | Hurt |
| Replace independent action-pair payoff outputs with a low-rank payoff approximation (ablation) | DCG · Relative overgeneralization predator-prey | sample efficiency | Helped |
| Use rank 1 rather than full payoff outputs (ablation) | DCG · Artificial decentralization predator-prey | sample efficiency | Helped |
| Add a privileged state-dependent bias function (ablation) | DCG · SMAC maps | performance | Mixed |
| Use DCG-S rather than QMIX | StarCraft II micromanagement · SMAC MMM2 | win rate | Helped |
| Use DCG-S rather than QMIX | StarCraft II micromanagement · SMAC 3s_vs_5z | win rate | Hurt |
| Add pairwise payoffs to VDN (ablation) | DCG versus VDN · SMAC maps | performance | No clear effect |
Closest papers by meaning
- Non-Linear Coordination Graphs (NeurIPS 2022 · similarity 0.90)
- Self-Organized Polynomial-Time Coordination Graphs (ICML 2022 · similarity 0.90)
- Context-Aware Sparse Deep Coordination Graphs (ICLR 2022 · similarity 0.89)
- Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments (NeurIPS 2017 · similarity 0.88)
- Graph Diffusion for Robust Multi-Agent Coordination (ICML 2025 · similarity 0.88)
- HYGMA: Hypergraph Coordination Networks with Dynamic Grouping for Multi-Agent Reinforcement Learning (ICML 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.